Is the Slipage in the SQL2K8 Release Date a Big Deal?

A few weeks ago we annouced an adjustment to the release date for SQL2K8. Many blog postings made a big deal about this. There were several consipiracy theories floating around. I'm not really sure why.

If you spent more than a minute and thought about it the slippage isn't that big of a deal at all. SQL Server is a huge product. Anyone who has worked on a big product knows about the complexity of shipping it. The bottom line is we won't ship until we believe (until we know) the product is ready. People should focus on that positive.

Just think about the test passes we run. We support 10+ languages running on WinXP, Vista, Win2K3, Win2K8, running on x86, x64, and IA64 platforms. We support cluster and non-cluster. We support upgrade from SQL2K, SQL2K5, and earlier SQL2K8 CTPs. That's one heck of a test matrix.

We also have requirements about production deployments of pre-released builds (both internal and external to Microsoft). We have numerous teams that need to do test passes (Analysis Services, Reporting Services, Integration Services, Replication, Data Programmability and Protocols, Manageability, Database Engine, etc). Each team has to sign-off on the build.

Running a full tests pass consists of 100,000's of tests requiring 1,000's of machine hours. We have long haul stress tests which take weeks to run. Every test failure is investigated to determine if it's an environmental issue, test issue, or product issue. All of this is done to ensure we release the best possible product.

One of the sayings around MS is that people will forget if you slip the release date a few weeks or months. But they'll never forget if you release a crappy product.

Rather than ridicule you should applaud us for having the guts to hold the product until it's ready. Trust me it's not a fun or easy conversation to have. After all when you make the decision to deploy SQL Server you want to have confidence in your decision. We know your reputation and in some cases your job or company is on the line.

So be different. Stand up and appluad us for doing the right thing for our customers. That will get you a lot more attention than taking the easy route of poking fun at us!

Comments (19)
  1. jeff.s says:

    I’ll be the first to say, here, thanks for holding it back. I know it’s a hard decision and a PR nightmare (albeit temporary). I think many developers have no real concept of both the complexity and the installed base of products from Microsoft. As a consultant, one thing I never want to say is "use the previous version" of a product because it was released prematurely. Ignore the naysayers, or just view them as proof people are excited about SQL Server 2008 🙂

    Keep up the good work

  2. Dave says:

    An enterprise class DBMS is something you don’t want to release if it’s not ready.

    Unfortunately (or not) people accept that there might be issues with a desktop (or even (heaven forbid) server) operating system and know it’ll get resolved in due course.

    But their database server needs to work correctly out of the box, and stay that way.  

    Database servers are systems you upgrade when you _have_ to, not just because the vendor has put out the new version.

    Sure it’ll put crimps in some peoples (myself included) project plans because they want to make use of upcoming features for their projects, but if you really need it right now, that’s what TAP is for.

    I am happy to wait…

  3. JamesNT says:


    You did the right thing.  There are people like me who know you did the right thing and we do applaud you for it.

    It’s important to remember that the people who are talking crap about you MAKE THEIR LIVING TALKING CRAP ABOUT MICROSOFT.  

    Mary Jo Foley, anyone?

    Ignore them.

    Microsoft is number one and will remain number one because you make the right decisions.


  4. davidacoder says:

    Give me a break. While you did the right thing to not release it if it is not ready, there is plenty to be critical of with this whole situation. I’ll just point out two:

    1. You can whine all day about the complexities of developing a large project like SQL Server, with the large test matrix etc. But, you know what? There are other teams that handle that well and ship on time with every single release. Office, anyone? They surely have a more complex test matrix than you guys have, and they deliver on time every time. Just because something is complex there is no reason to give, you still can manage that, as other teams succesfully do. It just appears that this kind of skill is not very strong within the SQL group. Some more hints of that? I saw a lot of people that were in charge of the whole WinFS mess resurface again in the SQL group. To me this whole thing simply is a sign that you do not have the leadership team in place in SQL that is capable of managing such a large and complex test matrix, and you don’t seem to have a culture of not simply throwing out those managers that have caused one of the most embarrasing delays of all Microsoft history (those in charge of WinFS), but instead keep them in influential positions.

    2. Even if this could not have been avoided, there is a HUGE problem with your timing and information policy. If I see things correctly, you informed the public end of January that you had a slip in the schedule, just four weeks before the previously anounced RTM date. And you essentially anounced a minimally six month slip. or something like that. I really ask myself the following: When did you realise that you would need more time? If this only became clear end of January, four weeks before your RTM date, then this would be sheer incompetence. You can tell me a lot about how complex SQL is, but any team that finds out four weeks before ship date that actually it needs six more months is seriously messed up and management obviously doesn’t know what is going on. The other option is that you knew much earlier. Why on earth did you not inform people about it then earlier?!? You must know that partners etc take your time lines into account in their own planning, and the minimal responsibility on your part to me seems to keep them updated about slips at the first moment you become aware of it.

    So, a delay is better than shipping crap, but this delay is in now way something to be proud of for the SQL team, but rather shows some serious problems in the management of the group.

    And finally, I thought the way the slip was anounced was some serious case of bad PR-style talk ("good news, hurray, lets cheer, we slip!") and this post is exactly the same kind of tone ("cheer us! we messed up! be thankful!"). I am just always remined of the communication from the WinFS team, every single slip from them was accompanied with great enthusiasm in their PR style blogging, and boy was that off.

  5. Hi Dan,

    I agree with most of the others here – in fact, I’d rather see the release of Katmai slip even further, even to 2009 if need be, than see an unfinished product hit the market.

    I don’t think I’ve written anything, in public or in private, that can be interpreted as chastising you for this delay. And if I did, then please accept my apologies for not expressing my thoughts clearly enough.

    However, I did criticize, and will continue to do so, the lunacy of having an "official launch event" months before the RTM. This was true with the original schedule, and has become even more true today. "Heroes happen here", eh? Yeah right – here but not now!

    I know that this is a PR decision that the dev folks have little -if any!- influence on. (And I can only begin to imagine the pressure the PR department puts on you folks to not delay RTM, so once more kudos for not budging but withholding the product until it’s good). But if you read the reaction posted by davidacoder (esp. point 2), you’ll see what misunderstandings have been caused by this socalled launch event. I expect NASA to launch several months before arrival, not Microsoft!

    Seriously – SQL Server 2008 should never have been included in the february launch event, not even at the original RTM schedule. And after this slip, the only sensible thing to do is to take SQL Server 2008 out of it, and reschedule a seperate launch event for Katmai.

  6. No one cares all that much if the product slips.  It’s a huge, complex product, and slips are totally understandable (perhaps even expected).  Personally, I’m with Jeff S. — I’m -thankful- that the product is slipping a bit, as I’ll have more time to learn it in depth before I actually need to use it.

    The comment by davidacoder hits the problem with the slip announcement perfectly.  The issue is not the slip itself, but rather the way that it was announced.  What’s up with the marketing back talk and exceptionally strange, circular wording?  Why not just say, "sorry, but the product will be a bit late."  That would have generated a lot less buzz.  There is no reason to try to sugarcoat this stuff.

    I think you should read Joel Spolsky’s post on the topic:

  7. Brian says:

    Microsoft ship something on time? Never. We don’t forget, we just expect delays.

    I’d rather have it late because, I’d rather not have a ton of bugs to have to work around. MS Office has plenty of that already.

    If MS wanted to look like they could produce a product on time, they would avoid creating buzz around a product before it’s ready to release. Then, when it is ready, they could announce that the product is ready and worth while because of all the new features. The current model appears to be to announce a product with a ton of new features in order to create buzz and then cut the bulk of the features in the name of releasing the product on time.

  8. I have been following Dan Jones’s blog on " Launch_T-1 , Launch_T-3 & Launch_T-5 " series about SQL

  9. I have been following Dan Jones’s blog on " Launch_T-1 , Launch_T-3 & Launch_T-5 " series about SQL

  10. Craig Manchester UK says:


    I don’t understand why some idiots decide to give you a hard time about this.  It’s not like we can’t survive a bit longer without this product. people like davidacoder should really think about getting a life and be thankful that he’s able to make his living from the great tools that you make for people like him.


    just release it when your ready 🙂

  11. Been reading Dan Jones blog post regarding the delay to the launch of SQL Server 2008, and I think he

  12. Paul Z says:

    Dan. Love your blog, informative.

    Think you’re missing something (humbly) and to see people post comments like ‘those idiot’s and so on puts a smile on my face.

    First, I was from development (at one point in my life). There is no way that you just found out 4 weeks in advance you were going to miss the GA no matter how complex the product is. Heck, a code freeze is typically months and months in advance. So unless you found a MAJOR bug — and I mean major, if you recall in SS05 they moved out MIRROR to SP1 cause they obviously found something…so….

    Now all that said, I admire pulling for quality, we need more of that. HOWEVER, how does marketing go along with the Heros Happen Here launch of SS08 without one of the heros? Just odd. I get business, I get the separation of development and marketing, but this was strange. Unfort’ly this is life in a tri-glued product set (which is good and bad).

    OK….but here’s another point….if you purchased SS05 under Software Assurance you were told you would have 36 month releases — that is the term of SA. That gives me a free upgrade. It seems to me you just created 6 months of folks that thought they would get a free upgrade to SS08 that won’t — unless they pay again for SA. Now if you paid 25% a year for your bill for upgrae protection and you got nothing well. On the other hand, you folks have delivered new features in your SPs so I admire that. So folks that comment on ‘idiots’ may be good at administering databases (and I was one of those too), but aren’t ‘up to par’ on procurement and TCO. Upgrade protection is a major revenue stream for MS, and Oracle’s Chuck Phillips tells investment banks that is their major thing they care about. So it does make you wonder…this is 8 years of SQL Server with no pay-off for SA.

    And finally, I feel sorry for you folks. Quite honestly, for SS05 everyone bitchced it was too long. Then for SS08 everyone started saying it was too short. Now everyone gets mad that it slips. Odd indeed — it’s like an argument with my mother (who moved the goal posts I think it goes).

    Bottom line, if MS is going to make upgrade protection contracts via SA 3 year terms, something should get delivered, especially when they openly promise they won’t slip an SA term. So can you offer free upgrades for those expired SA contraacts and then I’m not sure who can bitch…but someone will (likely Microsoft).

  13. Hey Paul,

    Thanks for the comments. I don’t have a comment on how marketing handled/communicated the slip in the release date. I’ve never been in marketing (nor have I ever played a marketing guy on TV) so I don’t really know what it’s like to manage through the complexity of releasing a product like SQL Server. I do know those guys have an entire world out there watching them ready to pounce on every misstep; certainly not an easy position to be in. The SQL2K5 campaign, the one with the batteries, was a very successful campaign so they get a ton of credit for that. I don’t know what they’ve cooked up for the SQL2K8 campaign; we’ll have to wait and see.

    I’m not an expert on SA, but since SQL2K5 released in October of 2005 so long as SQL2K8 ships by October of 2008 customers should reap the benefit. I might be missing something so feel free to educate me (and the readers of the blog) on the intricacies of SA.



  14. Paul says:

    Dan — I follow you as well, nice to see the intellecutal capital on this page.

    You are right about SS05, it RTM’ed Oct 2005. Sorry about that, I should have clarified more it’s really about when you purchased the SA contract and when it exprires and what you upgraded to.

    For those of you that don’t know, MS changed all their upgrade schemes (that wasn’t used in a bad way) to Software Assurance (SA) a number of years ago. So off went all the UPs, Volume Upgrades Competitive Uptradem License Upgrades, and so on. Well smart from consolidation, though the new SA replacements was more expensive in every case. In addition, it extended the lock-in term from 2 to 3 years. It originally was coming with something called GET CURRENT, in other words, at the time, if you didn’t move to the most current version of your product, you had to get current and then you could join SA; so SA, when you upgrade you pay full price. They backed away from that. Then SA didn’t get a lot of takers (obvious reasons). So MS has been steadily putting in features into it over the years. From support (albeit not worth it), to home use rights, training, and more. It’s gotten a lot of bette, but a lot of procurement officers and CIOs are still choked (kind of like when they pulled SQL Server from the CORE CAL). Anyway, I don’t want to bore everyone.. Upgrade in MS is handled by SA and support is a different manner. If the average support contract for an enterprise is 11% (it’s a tough one to come up with) and SA is 25% well you see where the money comes from. If you look at the $ of SQL Server (I love it!) then it isn’t there where the $$s roll in, it’s likely SA and CAL licensing. But Im straying beyond the intention..

    I feel bad for those guys too – got thrown under the bus! :). This is what happens when the products are so tightly tied — they stat to push off releases and mindshare generation and the company obviously decided that they wouldn’t let this happen.

    But you commentary did miss one point which you’re likely not in a position to answer anyway…you telling me they didn’t know until 4 weeks beforehand they were going to slip. That’s impossible in the Q/A dev process for software.

    I’m trying to be neutral here — hope I am coming across that way and I’m likely not. In the end, I think I agree with another commentator that no one will care.

    Yes the world was watching ready to pounce because MS has a history of slipping dates and that history costs people money.

  15. The marketing announcements are really more about timing than when they knew something. That’s part of the reason I’d have a hard time living in that world. If we know something why wouldn’t we tell everyone right away? They hold on to information (not withhold it) for specific reasons; usually to fit it into an overall communication plan. At least that’s theory.

    I think ther are two points: 1) The fact that we slipped the release date and 2) The communication of the slip. I should have been more clear that I was addressing the first one. I have no control over the second one.

  16. Paul Z says:

    I commented Dan on the SA here and for some reason it wasn’t showing up. I should have been more clear, the point is that there will more SA contracts expriing (depending when they are signed) that will not deliver a release upgrade for 3 years of paid service.

    But you’re point is right: living in that marketing world. I would have a hard time there.

    In the end – i say what I said, if you folks are within 3 years, everyone is saying too fast, yet everyone screamed at more or a slip. Somteimes you can’t win — it’s like an argument with my Greek mother: though that’s a different BLOG all together now isn’t it! 🙂

  17. Hi Paul,

    It’s strange that your comments aren’t showing up. I checked and I don’t have any pending comments. I think SQL2K5 was a bit of an anomaly on two fronts: 1) it set the wrong bar for release content and 2) it set the wrong bar for the interval between releases. Obviously there’s stuff going on internally that I can’t discuss, but let me say that Katmai is resetting the bar on both fronts. People can view this as a “dot” release all they want but when you step back and look at what’s been added to the product it’s a solid release that can stand on its own. As for the timing, it “feels” really fast for the product group as well. For some of us we’ve felt like we’ve been in ship mode for 2 years. Yikes! I’m confident we’ll get used to the rhythm and the cycle. Whenever I learn a new Green Day or Godsmack song on the drums it always feels super fast at first. But about the 5th or 6th time it starts feeling comfortable. I just hope it doesn’t take that many releases to start feeling comfortable with the release rhythm.



  18. Paul Z says:

    It could be my fat fingers why they aren’t showing up.

    I will tell you that SS08 is a release and all that. I’m not sure who can say it isn’t. Compression, workload management, and more. Kudos to your team.

    As for Green Day –be careful — love them, but all their music sounds the same.

    Now what’s all this talk about nothing past SP2 for SS05 — if a client is on SS05 and made that move and will stay there for a while, they really left alone on that?

Comments are closed.

Skip to main content